Abstract
Several researchers have called for investigation of the psychometric properties of widely used attributional measures. Models involving multiple indicators of a latent (unmeasured) variable provide a valuable approach to this type of problem. A strategy that gives evidence about the relative merit of several attributional measures is illustrated. In general it appears that the best item of those tested is one asking for ratings of the cause of an event on a scale from dispositional to situational. A question about predicted recurrence of the event serves almost as well, while questions about the actor's freedom in, and responsibility for, the event are substantially less adequate measures of an underlying attributional construct.
Get full access to this article
View all access options for this article.
